"<Elbert E Silbaugh>" <[log in to unmask]> wrote:
>Few (if any) of mine are that way. And I decided it was OK that they
>aren't because either the final expansion does not necessarily
>have to be the same size as all the previous ones; and/or
>
>The final block of a detail does not have to be full. Therefore,
>evenly divisible increments (as alluded to) may not always occur.
 
Elbert, you may be correct. I have not tested all scenarios. However, I might
be inclined, at this point, to disagree with you on the topic of "the final
expansion does not necessarily have to be the same size as the previous ones;".
 
Unfortunately, after I sent the message, I thought it would have been nice to
give an example of what I was referring to. So if you will stick with me for a
few more seconds. If we have a data set with a MaxCapacity of 100, an Initial
capacity of 25, and an Incremental amount of 10; we will encounter capacities
of 35,45,55...95. It is at this point, that when DDX is asked to expand the
set (by 10) it can't because that would put it over the MaxCapacity limit.
For which I kindly call a "stalled DDX". DDX was not designed to look at the
current capacity and choose a smaller incremental amount to keep it from going
over the MaxCapacity limit. The final expansion IS the same as the previous
ones...at least in my testing.
 
Hope this helps, because if you have the above scenario, I believe you will
run into the problem. This is a simple example, and does not take into account
the blocking factor multiple, as Ken Paul of ADAGER points out.
 
Best regards,
Steve Patrick
HP Database Expert Center